-=>Quoting Scott Mcnay to Darryl Gregorash <=-
SM>*** Darryl Gregorash wrote in a message to Bryan Schwartz:
DG>That was tongue-in-cheek, of course.. it can't be done,
DG>particularly if you have forward mem references whose exact
DG>location you cannot predict ahead of time.
SM>Of course you can do it; how do you think assemblers do it?
SM>You are correct on that specific case, though; he'd have to
SM>use symbolic names for relocatable constants, and resolve
SM>them after making changes.
SM>It's been a looong time since I did code by hand.  ;)
Nineteen-hundred-and-seventy-seven was that last time I bothered with
that. And _yes you can_ do it all.
Back then memonics were nice, but useless. You 'knew' what each
op-code did, and how many bytes followed for a particular referencing
mode. Granted, things were less complicated then. The 80xxx
instruction set is more complex, but would not be impossible.
Some collegues of mine wrote entire games this way. They did so even
though there were assemblers available. Were they 'real' programmers
or what! [grin]
But then, what else can you do when you are on a ship for sex
months...
SM>--Scott.
___ Blue Wave/QWK v2.12
--- Maximus/2 3.01
---------------